Pathway Analysis : An Introduction

Pathway Analysis : An Introduction

Experiments Literature and other KB Data Knowledge Structure in Data through statistics Structure in Knowledge through GO and other Ontologies Gain insight into Data Pathway Analysis

Why Pathway Analysis? Logical next step in any high-throughput experiment Treat samples Collect mrna Label Hybridize Scan Normalize Select differentially regulated genes Understand the biological phenomena involved Microarray Experiment High-throughput experiments per se do not produce biological findings Genes do not work alone, but in an intricate network of interactions Helps interpret the data in the context of biological processes, pathways and networks Global perspective on the data and problem at hand

Trends in Bioinformatics Sequence Comparison Today Functional Comparison Tomorrow Pathway Discovery Bridge to the Future and fully understanding of molecular basis of disease

Remember everything is a relationship (connected) what we are trying to do here is find that relationship (connection)

What do we get out of PA? In-depth and contextualized findings to help understand the mechanisms of disease in question Identification of genes and proteins associated with the etiology of a specific disease Prediction of drug targets Understand how to intervene therapeutically in disease processes Conduct targeted literature searches

What do we get out of PA? cont Data integration: integrate diverse biological information Scientific literature, knowledge databases Genome sequences Protein sequences, motifs and structures Functional discovery: assign function to genes Only 5% of known genes have assigned functions Without understanding the function, no drug discovery can be done

Werner T. Curr Opion Biotechnology 2008

Available Tools for Pathway Analysis (non-exhaustive list) GeneGo/MetaCore (www.genego.com) Ingenuity Pathway Analysis (www.ingenuity.com) Pathway Studio (www. ariadnegenomics.com) GenMAPP (www. genmapp.com) WikiPathways (www. wikipathways.org) cpath (cbio.mskcc.org/cpath) BioCyc (www.biocyc.org) Pubgene (www.pubgene.org) PANTHER (www. pantherdb.org) WebGestalt (bioinfo.vanderbilt.edu/webgestalt/) ToppGene Suite(/toppgene.cchmc.org/) DAVID (david.abcc.ncifcrf.gov/) Pathway Painter(pathway.painter.gsa-online.de/)

Available Databases (non-exhaustive list)

Why Pathway Analysis Software? A learning tool Study a group of gene products. A data analysis tool. Which pathways are particularly affected? What disease has similar biomarkers? A hypothesis generation tool Can provide insight into mechanisms of regulation of your genes. Which is the likely causative agent for the observed changes? What is likely to happen as a result of these changes? Suggest effects of gene knock-in or knock-outs. Suggest side-effects of drugs. Can highlight new phenomena that needs further investigation. What does the program not explain?

Caveat or how far the tools will take you in your quest for knowledge Tools are new Databases always evolving New Discoveries happen all the time

Caveats : Application used SNPs which showed association with T2D (Po0.003) were included in this study and were mapped backed to regions on the genome, and the predicted candidate genes were used for analysis.the top-10 ranking KEGG pathways per method are shown. Elbers et. al 2009

Caveats : Pathway DB used SNPs which showed association with T2D (Po0.003) were included in this study and were mapped backed to regions on the genome and the predicted candidate genes were used for analysis. The highest 10 ranking pathways per method are shown for Webgestalt BioCarta and PANTHER. Elbers et. al 2009

Caveats: Why Use of different databases Eg. KEGG, BioCarta, Properietary Use of different updates Use of different database updates Use of different statistical tests Use of different definitions/classification Ex. Some use inflammation while in others pathway is divided into inflammation related pathways like Jak-STAT signaling and cytokine-cytokine receptor interaction pathways. While some use hybrid models like GO hybrid (IPA) and others use GO (Metacore)

Biological Pathway Building Process Viswanathan G, et al. PLoS2008

Stages in Pathway Analysis 1 st Stage Analysis Data Driven Objective (DDO) Used mainly in determining relationship information of genes or proteins identified in a specific experiment (e.g. microarray study) Focused 2 nd Stage Analysis Knowledge Driven Objective (KDO) Used mainly in developing a comprehensive pathway knowledge base for a particular domain of interest (e.g. cell type, disease, system) Intergration Repeat 1 st Stage after generating new leads and hypothesis

Basic Concepts Node Symbolizes a list of, for example, genes. This is essentially a one-dimensional representation of the data Pathway Linked list of interconnected nodes. This is essentially a two-dimensional representation of the data Network A network of cellular functions and regulations involving interconnected pathways This is essentially a multi-dimensional representation of the data

Pathway Creation Algorithms in MetaCore Analyze Network: Creates a list of possible networks, ranked according to how many objects in the network correspond to the user's list of genes, how many nodes are in the network, how many nodes are in each smaller network. Analyze Networks (Transcription Factors): For every transcription factor (TF) with direct target(s) in the root list, this algorithm generates a sub-network consisting of all shortest paths to this TF from the closest receptor with direct ligand(s) in the root list. Shortest paths: Uses Dijkstra s shortest paths algorithm to find the shortest directed paths between the selected objects. Self regulation :Finds the shortest directed paths containing transcription factors between the selected objects

Direct interactions: Draws direct interactions between selected objects. No additional objects are added to the network Auto expand : Draws sub-networks around the selected objects, stopping the expansion when the sub-networks intersect. Transcription regulation : Generates sub-networks centered on transcription factors. Sub-networks are ranked by a P-value and interpreted in terms of Gene Ontology. Analyze network (receptors) :For every receptor with direct ligand(s) in the root list, this algorithm generates a sub-network consisting of all shortest paths from that receptor to the closest TF with direct target(s) in the root list.

An Example to illustrate the Stages in Pathway Analysis 1 st Stage Analysis Data Driven Objective (DDO) Used mainly in determining relationship information of genes or proteins identified in a specific experiment (e.g. microarray study) Focused topic of interest 2 nd Stage Analysis Knowledge Driven Objective (KDO) Used mainly in developing a comprehensive pathway knowledge base for a particular domain of interest (e.g. cell type, disease, system) Broad topic of interest Repeat 1 st Stage after generating new leads and hypothesis

Example MicroRNA network interactions in REH/MSC cells

mirna s are 22-nucleotide non-coding RNAs that regulate gene expression through base pairing with target mrna Endogenous Regulatory Functions Invertebrates developmental timing neuronal differentiation cell proliferation, growth control, programmed cell death Mammals embryogenesis stem cell maintenance hematopoietic cell differentiation brain development

Background Experiments were performed to analyze the effect of low oxygen conditions and the interaction with the microenvironment in the expression pattern of microrna s in REH cells. In this project we look at the possible interactions between the measured microrna s with other molecules related to the pathogenesis of lymphocytic leukemia. In this initial stage we plan to put this complex system of microrna interactions in context with its surrounding interactions. It was discerned from the current analysis that the anti-apoptotic action of microrna 21 may be due to its interaction with Bcl-2 and MCL-1 in MSC cells. However, this needs to be further explored. In this initial report we looked at all the tested micrornas in the context of its associated biological networks.

Experimental Details Briefly, REH cells were cultured alone or co-cultured with N.BM MSC (or H-Tert immortalized MCS) for 24 h and 48 h under different po2 conditions. At the end of 24 h and 48 h REH cells and MCS cells were sorted by Flow Cytometry and each cell population was lysed with TRIZOL to extract total RNA separately. MicroRNA assay details: The biotin-labeled cdna targets are prepared by a simple reverse transcription into first strand cdna. Total RNA is primed for reverse transcription by a random Octomer conjugated with two biotins and a 5 poly (A) tail. This procedure results in an equal copy number of biotin cdna target to the templates of mirna (see figure on left). Two 40 mer oligo probes, one for the mature mirna and the other for precursor oligo, were designed from the sense strand of both arms of the hairpin structure of the microrna precursor sequence collected from the Sanger Database. The oligo probes were modified at the 5 end with Amine-C6 linker and ordered from Integrated DNA technology (IDT) (Coralville, IA, USA) at 50 or 100 μm stock concentration in H2O (see figure on left). REH: pre-b Acute lymphoblastic leukemia cell line MSC: Mesenchymal stroma cell line

1: 1 st Stage Analysis NETWORK RELATIONSHIPS BASED ON THE EXPERIMENTAL INPUT

microrna interaction partners Hub interactions

microrna 21 and its relationship to Bcl-2 Why BCL-2 Inhibits BCL-2 activity

microrna 21, Mcl-1, interactions Why Mcl-1 Mir21 indirectly interacts with Mcl-1

2: 1 st Stage Analysis GO ENRICHMENT FOR THE MEASURED MICRORNA

GO enrichment for the measured microrna # Process % p-value 1 release of cytochrome c from mitochondria 30.77 6.35E-09 2 general transcription from RNA polymerase II 23.08 7.7E-09 promoter 3 cell death 61.54 2.47E-08 4 death 61.54 2.67E-08 5 regulation of gene expression 92.31 3.29E-08 6 apoptotic mitochondrial changes 30.77 3.43E-08 7 multi-organism process 61.54 1.09E-07 8 regulation of macromolecule metabolic process 92.31 1.78E-07 9 regulation of metabolic process 92.31 2.81E-07 10 regulation of macromolecule biosynthetic 84.62 7.95E-07 process 11 regulation of cellular biosynthetic process 84.62 1.03E-06 12 regulation of biosynthetic process 84.62 1.06E-06

3: 1 st Stage Analysis DISEASE REPRESENTATION FOR THE MEASURED MICRORNA

Network-disease associations # Disease % p-value 1 Lymphoma, Intermediate-Grade 23.81 5.59E-09 2 Lymphoma, Diffuse 26.19 9.94E-09 3 Lymphoma, Small-Cell 23.81 5.7E-08 4 Lymphoma, Mantle-Cell 21.43 7.72E-08 5 Lymphoma, Small Cleaved-Cell, Diffuse 21.43 7.72E-08 6 Decompression Sickness 7.14 9.71E-08 7 Barotrauma 7.14 9.71E-08 8 Vaccinia 14.29 2.05E-07 9 Poxviridae Infections 14.29 4.19E-07 10 Leukemia, B-Cell 14.29 8.59E-07 11 Lymphoma, B-Cell 26.19 9.68E-07 12 Lymphoma, High-Grade 14.29 3.27E-06

4: 1 st Stage Analysis POSSIBLE THERAPEUTIC TARGETS AND OTHER INTERACTIONS FOR THE MICRORNA NETWORK

No direct activators/ inhibitors for mir21 However, indirect activators /inhibitors for mir21

Questions What new things have we learned? What type of things should we expect to learn in general? What new experiments do these suggest? How can this new knowledge be exploited?

Stages in Pathway Analysis 1 st Stage Analysis Data Driven Objective (DDO) Used mainly in determining relationship information of genes or proteins identified in a specific experiment (e.g. microarray study) Focused topic of interest 2 nd Stage Analysis Knowledge Driven Objective (KDO) Used mainly in developing a comprehensive pathway knowledge base for a particular domain of interest (e.g. cell type, disease, system) Broad topic of interest Repeat 1 st Stage after generating new leads and hypothesis On and on we go

Demo Search and Explore Building pathways Analysis Canonical Pathways

A caveat Not every gene belongs to a pathway in the database Pathways generated are a statistical probability rather than a biological certainty Context context context it matters a lot in pathway analysis Findings need to be proved experimentally

DB Similarity Search Chemical properties, QSAR, etc Compound Predicted Metabolites Similar DB Compounds Putting it together Network Analysis Interconnecting compound/metabolites with their predicted targets and other associated network objects. Experimental Data (Optional) Import and overlay any experimental data Lists of Targets (genes/proteins) for the input molecule and for metabolites Based on the known targets for the Similar Compounds Functional Enrichment Analysis Based on the Predicted Targets: -Drug target networks -Toxicity networks -Canonical Pathway Maps -Process networks -GO processes -Disease Networks Integrative Systems-Level Summary of Predicted Primary and Secondary Effects of a query Compound